## 2.  Country Coverage and Characteristics of Banking Sector

## Source details

**Canonical URL:** [2.  Country Coverage and Characteristics of Banking Sector](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2009/_wp09273.pdf)

## Other formats

- [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2009/_wp09273.pdf.md)
- [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2009/_wp09273.pdf.json)

---

### I. Introduction
- Study question: Which factors determine whether foreign banks outperform or underperform domestic banks in host developing countries?
- Sample and period:
  - Large group of countries over the period 1999-2006.
  - Panel data of foreign and domestic commercial banks.
- Motivation: Among ~35 studies reviewed, 15 found foreign banks perform better on all measures, 9 found worse or no difference on all measures, others mixed.
- Key hypothesis: The advantage or disadvantage of being foreign depends on home-country characteristics, host-country conditions, distance (geographical/cultural/institutional), and bank-specific characteristics (size, funding, age).

### II. Theoretical predictions and related literature
- Predicted advantages of foreign banks:
  - Efficiency gains from servicing multi-country clients and spreading best practices.
  - Better risk diversification and access to parent liquidity, potentially lowering funding costs.
  - Scale advantages enabling sophisticated risk models.
- Predicted disadvantages of foreign banks:
  - Less local information, initial informational disadvantage.
  - Potential discrimination by host governments/customers.
  - Diseconomies from operating and monitoring across distances and cultural/institutional differences.
- Factors emphasized:
  - Home country development and competition conveying technical/regulatory advantages.
  - Host country banking efficiency, financial development and competition affecting entry and profit opportunities.
  - Distance (geographical, cultural, institutional) increasing information asymmetries and monitoring costs.
  - Bank-level factors: size, funding and asset mixes, time present in host country (learning/adaptation).

### III. Data and empirical methodology
- Database and definitions:
  - Ownership and balance-sheet information for banks in all developing countries over 1995-2006; coverage accounts for roughly 90 percent or more of banking system assets in each country in the latter period.
  - Bank is foreign owned if 50 percent or more of shares owned by foreigners; home country assigned by country with highest percentage of foreign shares; direct ownership used, except where direct owner is a tax vehicle.
  - Ownership information obtained primarily from Bankscope, complemented by banks’/parents’ websites, regulators, stock exchanges, SEC Form F-20, country experts; ownership info obtained for almost 95 percent of banks for their active periods.
- Sample selection to limit endogeneity from entry choice:
  - Include countries with at least 3 foreign banks active over entire sample period and at least 3 domestic banks active over entire sample period.
  - Using 1999-2006 yields a subsample of 51 countries (results robust to different samples).
- Key variables (selected):
  - Home/host development: gdpcap; dummy developing; dummy low for host lower income (World Bank 2006).
  - Financial development (findev): M2 divided by GDP.
  - Competition: Panzar Rosse H-statistic (H) with interpretation H<0 monopoly, H=1 perfect competition, 0<H<1 monopolistic competition; H-statistic available for 50 countries and is not time-varying.
  - Distance proxies: samereg, comlang, instfam (based on KKM governance indicators range -2.5 to 2.5).
  - Bank-level: share (assets as % of domestic banking assets), deposit (deposits/liabilities), loan (loans/assets), leverage (equity/assets), public (majority government owned), problembank (exited within 4 years), age cutoff 8 years.
- Empirical model and estimation:
  - Dependent variable: profitability = profit before taxes divided by total assets.
  - Specification: model (1) ICT ICT ICT ... (text form: Π = α0 + α1 Own + β'X + γFOwn + ε), estimated by OLS with country-year fixed effects, robust standard errors clustered at host country level, weights equal to inverse of number of banks in host country, exclude first 2 years after bank entry/acquisition.
  - Controls: continuous share, loan, deposits; leverage; public; problembank.
  - Observations in main bank-level summaries: 7,923 (full sample), foreign-bank observations: 2,540 (except competition_home 2,281 and competition_host 1,685).

### IV. Empirical results
- Individual country regressions (51 countries):
  - Foreign banks perform better (ownership positive and significant) in 14 countries.
  - Foreign banks perform worse (ownership negative and significant) in 8 countries.
  - In the majority, no significant difference; within the insignificant group ownership positive in 14 and negative in 15.
  - Conclusion: Heterogeneity—being foreign can be asset, liability, or neutral depending on context.
- Home and host country characteristics (key findings):
  - Aggregate: Without heterogeneity, foreign ownership has no significant impact on profitability.
  - Home-country development:
    - Own*gdpcap_home positive and significant: foreign banks with parents in high-income countries outperform domestic banks.
    - Own*developing negative and significant: foreign banks from developing countries perform worse than domestic banks.
    - Result robust to controlling for foreign bank scale.
  - Competition:
    - Competition in home country (comp_home) does not affect performance.
    - Competition in host country (comp_host) matters: Own*comp_host negative and significant—when host-country competition is limited (lower H), foreign banks more likely to outperform domestic banks.
  - Illustrative economic magnitudes (examples from text):
    - A foreign bank from a high income country investing in the host country with lowest competition (Turkey) earns on average a profit before tax of 0.72 higher than a domestic bank (0.72 equals 44 percent of mean profitability).
    - The same bank in country with strongest competition (Costa Rica) earns on average 0.70 less than a domestic bank.
    - A foreign bank from a developing country earns on average 0.18 less compared to a domestic bank in host country with lowest competition and 1.60 less in host country with highest competition.
  - Bank-level controls:
    - Large banks more profitable than smaller banks.
    - Higher loan ratio associated with higher profitability.
    - Higher leverage (equity share) associated with higher profitability.
    - Public (majority government owned) domestic banks less profitable.
    - Problembank associated with lower profitability.
- Distance effects:
  - Geographical and cultural proximity:
    - Own*samereg positive and significant: home and host in same region → higher profitability for foreign bank.
    - Own*comlang positive and significant: shared language → higher profitability for foreign bank.
    - Both high-income and developing-country foreign banks benefit from geographical/cultural closeness.
  - Institutional familiarity:
    - Own*instfam not significant; institutional distance measured by KKM governance indicators did not show significant impact.
    - Robustness checks with World Bank Doing Business indicators also did not find institutional distance effects.
- Bank characteristics (size, funding, age):
  - Size and market power:
    - Large foreign banks significantly outperform small foreign banks and domestic banks (large foreign bank coefficient 0.812***).
    - Foreign banks with monopoly power in host country significantly outperform other groups (foreign bank with monop. power coefficient 0.777***).
  - Funding structure:
    - Foreign banks with many deposits outperform domestic banks and foreign banks with limited deposits (foreign bank many deposits coefficient 0.410*).
  - Asset structure:
    - No statistically significant differences across groups in loan intensity.
  - Age/time presence:
    - Old foreign banks (more than 8 years) have the best performance: profitability about 0.4 percentage points higher than others (overall average profitability is 1.6 percentage points).
    - Confirms adaptation/learning over time, though survivor bias may play a role.
- Synthesis:
  - Foreign banks from high-income home countries tend to be more profitable than domestic banks; from developing home countries they tend to be less profitable.
  - Entering host countries with limited competition increases foreign bank profitability.
  - Geographical and cultural proximity (same region, shared language) increase foreign bank profitability; institutional familiarity (as measured) not significant.
  - Bank size, time present (>8 years), and funding structure (local deposit base) are important determinants of higher relative profitability.

### V. Conclusions and implications
- Main conclusion: Heterogeneity among foreign banks and in home-host contexts explains conflicting results in prior literature; foreignness is neither universally an asset nor a liability.
- Key determinants of relative profitability of foreign banks:
  - Home-country development (high-income parent banks more advantageous).
  - Host-country competitiveness (less competition more favorable).
  - Geographical and cultural proximity between home and host.
  - Bank-specific characteristics: size, time in market (age), funding structures (local deposit base).
- Longer-run implication:
  - Advantages of large foreign banks may drive further international consolidation.
  - Rising economic similarity of some emerging markets to high-income countries, and the importance of geographical/cultural proximity, suggest banks from certain emerging markets may play an increasing role in cross-border banking within developing regions.

### Appendix: Variable Definitions and Sources (selected)
- Profitabality: Profit before tax divided by total assets. Source: Bankscope.
- Own: Dummy = 1 if 50 percent or more of shares owned by foreigners. Source: Claessens, Van Horen, Gurgarlan and Mercado (2008).
- Share: Assets of the bank divided by total assets in the banking system of the country. Source: Bankscope.
- Loan: Total loans divided by total assets of the bank. Source: Bankscope.
- Leverage: Total equity divided by total assets of the bank. Source: Bankscope.
- Deposit: Total deposits and short-term funding divided by total liabilities of the bank. Source: Bankscope.
- Public: Dummy = 1 if the bank is for 50 percent or more owned by the government. Source: Micco, Panizza and Yanez (2007).
- Problembank: Dummy = 1 if the bank exited the market within 4 years after entry. Source: Claessens, Van Horen, Gurganlar and Mercado (2008).
- Gdpcap: GDP per capita in current international $ in host or home country. Source: World Development Indicators.
- Comp: Panzar Rosse H-statistic as calculated by Claessens and Laeven (2004). Source: Claessens and Laeven (2004).
- Findev: M2 divided by GDP in the host country. Source: International Financial Statistics.
- Samereg: Dummy = 1 if home and host country share the same region. Source: World Bank.
- Comlang: Dummy = 1 if home and host country share same language. Source: CIA World Factbook (2005).
- Instfam: Dummy based on KKM governance indicators average absolute difference below median = 1. Source: Kaufmann, Kraay and Mastruzzi (2008).
- Businessfam: Same as instfam but using five Doing Business indicators. Source: Doing Business indicators.

*Source: _wp09273 - 2.  Country Coverage and Characteristics of Banking Sector.....................................25*

### 2.  Country Coverage and Characteristics of Banking Sector.....................................25

### 2.  Country Coverage and Characteristics of Banking Sector

### Section location
- Page 25: "Country Coverage and Characteristics of Banking Sector"

*Source: _wp09273 - 2.  Country Coverage and Characteristics of Banking Sector.....................................25*

### References...............................................................................................31

### _wp09273 - References...............................................................................................31

### I. Introduction
- Study question: Which factors determine whether foreign banks outperform or underperform domestic banks in host developing countries?
- Sample and period: large group of countries over the period 1999-2006; panel data of foreign and domestic commercial banks.
- Motivation: Literature shows ambiguous results: among ~35 studies reviewed, 15 found foreign banks perform better on all measures, 9 found worse or no difference on all measures, others mixed.
- Key hypothesis: The advantage or disadvantage of being foreign depends on home-country characteristics, host-country conditions, distance (geographical/cultural/institutional), and bank-specific characteristics (size, funding, age).

### II. Theoretical predictions and related literature
- Predicted advantages of foreign banks:
  - Efficiency gains from servicing multi-country clients and spreading best practices.
  - Better risk diversification and access to parent liquidity, potentially lowering funding costs.
  - Scale advantages enabling sophisticated risk models.
- Predicted disadvantages of foreign banks:
  - Less local information, initial informational disadvantage.
  - Potential discrimination by host governments/customers.
  - Diseconomies from operating and monitoring across distances and cultural/institutional differences.
- Factors emphasized in theory and prior empirical work:
  - Home country development and competition can convey technical/regulatory advantages.
  - Host country banking efficiency, financial development and competition affect ease of market entry and profit opportunities.
  - Distance (geographical, cultural, institutional) increases information asymmetries and management/monitoring costs, lowering foreign bank performance.
  - Bank-level factors: size (economies of scale), funding and asset mixes, and time present in host country (learning/adaptation).

### III. Data and empirical methodology
- Database:
  - Ownership and balance-sheet information for banks in all developing countries over 1995-2006; coverage accounts for roughly 90 percent or more of banking system assets in each country in the latter period.
  - Definition: bank is foreign owned if 50 percent or more of shares owned by foreigners; home country assigned by country with highest percentage of foreign shares; direct ownership used, except where direct owner is a tax vehicle.
  - Ownership information obtained primarily from Bankscope, complemented by banks’/parents’ websites, regulators, stock exchanges, SEC Form F-20, country experts; ownership info obtained for almost 95 percent of banks for their active periods.
- Sample selection to limit endogeneity from entry choice:
  - Include countries with at least 3 foreign banks active over entire sample period and at least 3 domestic banks active over entire sample period.
  - Using 1999-2006 yields a subsample of 51 countries (results robust to different samples).
- Host/home characteristics and variables:
  - Home/host development: GDP per capita (gdpcap), dummy developing (parent in developing country), dummy low for host lower income (World Bank 2006).
  - Financial development (findev): M2 divided by GDP.
  - Competition: Panzar Rosse H-statistic (H): H<0 monopoly, H=1 perfect competition, 0<H<1 monopolistic competition; H-statistic available for 50 countries and is not time-varying.
  - Distance proxies:
    - samereg = 1 if home and host in same World Bank region.
    - comlang = 1 if home and host share same language.
    - instfam = 1 if absolute difference of averaged KKM governance indicators (six dimensions) between host and home is smaller than median difference; KKM indicators range -2.5 to 2.5.
  - Bank-level variables:
    - share = bank assets as percentage of domestic banking sector assets.
    - deposit = ratio of deposits to liabilities.
    - loan = loan to asset ratio.
    - leverage = equity divided by assets.
    - public = 1 if domestic bank majority government owned.
    - problembank = 1 if bank exited market within four years after entry.
    - age categories: cutoff 8 years for old vs new; for other variables create dummies above/below median and interact with ownership to form four groups (small domestic, large domestic, small foreign, large foreign).
- Empirical model:
  - Dependent variable: profitability = profit before taxes divided by total assets.
  - Specification: ictictictictictict XFOwnOwnεγββα ++++=Π'* 1 210 (model (1) in text), estimated by OLS with country-year fixed effects, robust standard errors clustered at host country level, weights equal to inverse of number of banks in host country, exclude first 2 years after bank entry/acquisition.
  - Controls: continuous share, loan, deposits; leverage; public; problembank. Observations in main bank-level summaries: 7,923 (full sample), foreign-bank observations: 2,540 (except competition_home 2,281 and competition_host 1,685).

### IV. Empirical results
- Individual country regressions (model (1) with year fixed effects only):
  - Among 51 countries:
    - Foreign banks perform better than domestic banks (ownership positive and significant) in 14 countries.
    - Foreign banks perform worse (ownership negative and significant) in 8 countries.
    - In the majority, no significant difference; ownership positive in 14 and negative in 15 within the insignificant group.
  - Reinforces heterogeneity: being foreign can be asset, liability, or neutral depending on context.

- Home and host country characteristics (Table 5 main findings):
  - Aggregate: Without heterogeneity, foreign ownership has no significant impact on profitability.
  - Home-country development:
    - Own*gdpcap_home positive and significant (income effect); foreign banks with parents in high-income countries outperform domestic banks.
    - Own*developing negative and significant: foreign banks from developing countries perform worse than domestic banks.
    - Result robust to controlling for foreign bank scale.
  - Competition:
    - Competition in home country (comp_home) does not affect performance.
    - Competition in host country (comp_host) matters: Own*comp_host negative and significant—when host-country competition is limited (lower H), foreign banks more likely to outperform domestic banks.
  - Economic magnitudes (examples given in text):
    - A foreign bank from a high income country investing in the host country with lowest competition (Turkey) earns on average a profit before tax of 0.72 higher than a domestic bank (0.72 equals 44 percent of mean profitability).
    - The same bank in country with strongest competition (Costa Rica) earns on average 0.70 less than a domestic bank.
    - A foreign bank from a developing country earns on average 0.18 less compared to a domestic bank in host country with lowest competition and 1.60 less in host country with highest competition.
  - Bank-level controls consistently:
    - Large banks more profitable than smaller banks.
    - Higher loan ratio associated with higher profitability.
    - Higher leverage (equity share) associated with higher profitability.
    - Public (majority government owned) domestic banks less profitable.
    - Problembank associated with lower profitability.

- Distance effects (Table 6 main findings):
  - After controlling for home country income, geographical and cultural proximity matter:
    - Own*samereg positive and significant: home and host in same region → higher profitability for foreign bank.
    - Own*comlang positive and significant: shared language → higher profitability for foreign bank.
    - Both high-income and developing-country foreign banks benefit from geographical/cultural closeness.
  - Institutional familiarity:
    - Own*instfam not significant; institutional distance measured by KKM governance indicators did not show significant impact.
    - Robustness checks with World Bank Doing Business indicators also did not find institutional distance effects; interpretation: institutional distance may be easier for foreign banks to overcome compared to geographic/cultural distance or the indicators used may be too general.

- Bank characteristics (Table 7 main findings):
  - Size and market power:
    - Large foreign banks (large foreign bank dummy) significantly outperform small foreign banks and domestic banks (large foreign bank coefficient 0.812***).
    - Foreign banks with monopoly power in host country significantly outperform other groups (foreign bank with monop. power coefficient 0.777***).
  - Funding structure:
    - Foreign banks with many deposits outperform domestic banks and foreign banks with limited deposits (foreign bank many deposits coefficient 0.410*).
    - Suggests that foreign banks with strong local deposit bases compete more effectively.
  - Asset structure:
    - No statistically significant differences across groups in loan intensity; possible offsetting effects of superior risk-taking vs higher transaction costs.
  - Age/time presence:
    - Old foreign banks (more than 8 years) have the best performance: profitability about 0.4 percentage points higher than others (overall average profitability is 1.6 percentage points).
    - Confirms adaptation/learning over time, though survivor bias may play a role.

- Synthesis of empirical findings:
  - Foreign banks from high-income home countries tend to be more profitable than domestic banks; from developing home countries they tend to be less profitable.
  - Entering host countries with limited competition increases foreign bank profitability.
  - Geographical and cultural proximity (same region, shared language) increase foreign bank profitability; institutional familiarity (as measured) not significant.
  - Bank size, time present (>8 years), and funding structure (large deposit base) are important determinants of higher relative profitability.

### V. Conclusions and implications
- Main conclusion: Heterogeneity among foreign banks and in home-host contexts explains the conflicting results in prior literature; foreignness is neither universally an asset nor a liability.
- Key determinants of relative profitability of foreign banks:
  - Home-country development (high-income parent banks more advantageous).
  - Host-country competitiveness (less competition more favorable).
  - Geographical and cultural proximity between home and host.
  - Bank-specific characteristics: size, time in market (age), funding structures (local deposit base).
- Longer-run implication for global banking:
  - Advantages of large foreign banks may drive further international consolidation.
  - Rising economic similarity of some emerging markets to high-income countries, and the importance of geographical/cultural proximity, suggest banks from certain emerging markets may play an increasing role in cross-border banking within developing regions.

*Source: _wp09273 - References...............................................................................................31*

### Appendix Table 1. Variable Definitions and Sources

### Appendix Table 1. Variable Definitions and Sources

### Banklevel
- Profitabality  
  - Definition: Profit before tax divided by total assets  
  - Source: Bankscope

- Own  
  - Definition: Dummy which is one if 50 percent or more of the shares of the bank are owned by foreigners, zero otherwise.  
  - Source: Claessens, Van Horen, Gurgarlan and Mercado (2008)

- Share  
  - Definition: Size of the bank. Assets of the bank divided by total assets in the banking system of the country.  
  - Source: Bankscope

- Loan  
  - Definition: Total loans divided by total assets of the bank.  
  - Source: Bankscope

- Leverage  
  - Definition: Total equity divided by total assets of the bank.  
  - Source: Bankscope

- Deposit  
  - Definition: Total deposits and short-term funding divided by total liabilities of the bank.  
  - Source: Bankscope

- Public  
  - Definition: Dummy which is one if the bank is for 50 percent or more owned by the government, zero otherwise.  
  - Source: Micco, Panizza and Yanez (2007)

- Problembank  
  - Definition: Dummy which is one if the bank exited the market within 4 years after entry, zero otherwise.  
  - Source: Claessens, Van Horen, Gurganlar and Mercado (2008)

### Home and host characteristics
- Gdpcap  
  - Definition: GDP per capita in current international $ in host or home country.  
  - Source: World Development Indicators

- Comp  
  - Definition: Panzar Rosse (1987) H-statistic as calculated by Claessens and Laeven (2004).  
  - Source: Claessens and Laeven (2004)

- Findev  
  - Definition: M2 divided by GDP in the host country  
  - Source: International Financial Statistics

- Distance  
  - Definition: (no definition or source provided in table)

- Samereg  
  - Definition: Dummy which is one if home and host country share the same region, zero otherwise.  
  - Source: World Bank

- Comlang  
  - Definition: Dummy which is one if home and host country share same language, zero otherwise.  
  - Source: CIA World Factbook (2005)

- Instfam  
  - Definition: Dummy capturing whether home and host country are institutional distant or not. First the absolute difference between the quality of institutions of source and host countries, based on the simple average of the absolute difference of each of the six governance indicators, is calculated. When the difference is below the median difference the dummy has a value 1 if it is above it has a value zero.  
  - Source: Kaufmann, Kraay and Mastruzzi (2008)

- Businessfam  
  - Definition: Same as instfam but difference in quality of five doing business indicators (cost of registering property, legal rights index, credit information, investor protection index, cost of enforcing contracts).  
  - Source: Doing Business indicators

*Appendix Table 1. Variable Definitions and Sources*

---


_Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2009/_wp09273.pdf_